Nucleation Delay in Selective Atomic Layer Deposition: Density Functional Insights Coupled Numerical Nucleation Model

Abstract

Selective atomic layer deposition (S-ALD) is gaining significant attention for its pivotal role in the bottom-up manufacturing of semiconductors and catalysts. This study delves into the selective growth mechanisms of M(N t BU)(NEt 2 ) 3 (M = Nb and Ta) on Cu and SiO 2 substrates using density functional theory (DFT) calculations. We discovered that the nucleation delay on Cu substrates is primarily due to sluggish dissociation kinetics. Utilizing these insights, we propose an anisotropic growth model framed within microkinetics and numerical nucleus expansion approaches. Our model integrates defect-and diffusion-induced nucleation parameters, aligning closely with experimental growth curves and accurately quantifying ALD nucleation delays in the initial stages. Such a model builds a direct link between the observed nucleation delay and ALD experimental parameters from the atomic scale. Furthermore, this DFT-based nucleation model demonstrates a robust predictive capacity for nucleation delay cycles for a wide range of materials, aligning well with the experimental findings. This advancement offers a theoretical method for assessing the selectivity of S-ALD from the bottom up, enhancing our understanding and control of these critical processes.